Recently, a technique regarding a large size database management system, such as a large amount of structured or unstructured data set which is so large to be collected, stored, managed, and analyzed using a database management tool of the related art, and big data processing which extracts a value from the data and analyzes a result thereof, has been developed.
When a database server performs a database operation, the operation is performed in a memory to shorten an I/O time. For example, when the database server changes data stored in a permanent storage medium, the database server loads data from the permanent storage medium, stores a loaded data in a cache memory, changes an object stored in the cache memory, and then reflects the changed data to the permanent storage medium.
In this case, since the data which is stored in the cache memory is later used for a database operation, the data may be stored without being deleted.
The database server may use a shared memory as a cache memory. The shared memory refers to a memory which may be used by a plurality of processors and the database server uses the shared memory as a cache memory, thereby improving data management efficiency.
Therefore, in order to use the shared memory as a cache memory, a technique which efficiently manages the shared memory needs to be studied.